Improving the state of Guile in Fedora: how to get involved?

Project Discussion
,
1

The state of GNU Guile Scheme in Fedora is not terrible, but it’s also not great. Bugs often seem to languish on RHBZ (notably at present: the interpreter itself is now a couple of releases behind and a dependency for standard library functionality isn’t available for Guile 3.0), the installed file paths are sufficiently non-standard to make compiling almost any third-party library require special configure invocations, there is a dearth of libraries available (a common one, guile-lib, used to be packaged but seemed not to make it to F29+) and Guile support was removed from GDB in F38.[1]

I know that Guile is something of a niche environment, so I’ve just kind of put up with the fact that it’s not going to be high on anyone’s list of priorities. But A serious suggestion for Fedora contributors passionate about privacy made me realise that not only is helping out a theoretical possibility, but that actually almost all the above issues are eminently tackleable; I even have a pile of specs that I’ve been pushing to Copr for a handful of libraries.

The only thing is, I’m not really sure how to go from “I have the ability and the motivation to fix these things” to actually getting those fixes into Fedora.[2] I was wondering if maybe someone could give me some guidance about how to get started?

  1. This was an extra sad discovery as writing instrumentation in Guile Scheme is a delight and it means I don’t get to enjoy my upstream contributions to GDB for Guile OOTB in Fedora (the horror!) ↩︎

  2. Besides new packages, I’m aware there are guides for those. ↩︎

2

I think we used to have some better documentation on this (on helping with an existing package, not guile specifically), but it seems to have bit-rotted a bit. Do the following docs help?

1 Like
3

I’m interested in eventually forming a Guile SIG, given I’m interested in Guix and there’s so many missing dependencies … and all the Guile packages throw a lot of rpmlint warnings, so work is needed on standardizing the directory structure and reconciling rpmlint to that!

Can’t promise too much time for this, but that is precisely why it can’t be a one person effort anyway. I last touched this a couple of years ago, remember making use of someone’s COPR so if that is you, thank you!

(Got guix to compile and it immediately crashed the desktop session on start, while the prebuilt binary worked, and I have not had time to look deeper)

2 Likes
4

Forgot to say, I am a sponsor and happy to work with anyone interested to get them qualified as packagers

2 Likes
5

Oh, right, the documentation. I really should have had a search before bothering people with a new topic (although, if I’m honest, I’m glad I didn’t as having some human feedback has been encouraging). Thank you for spoon feeding me a bit here :slightly_smiling_face:

It seems like the roadmap for me might be to try and get some of my existing packages into Fedora proper to learn any skills I might be missing and demonstrate that I’m capable of doing the work, applying for packager sponsorship (thank you @salimma for your offer, btw!) and then applying for co-maintainership of floundering packages. Does this sound right?

1 Like
6

it’s no problem — as I said, the docs are a bit bit-rotted. And I agree that it’s nice to have some human connections to get started on something!

1 Like